CN104052709A - Method and apparatus for performing soft demapping - Google Patents

Abstract

The invention provides a method and an apparatus for performing soft demapping. The apparatus includes a pre-processing unit to pre-process a reception signal obtained from a symbol representing bits; a candidate selection unit to select two candidates from among constellation points included in a constellation for each of the bits; a distance calculation unit to calculate a Euclidean distance between the reception signal and the two candidates; and a log-likelihood ratio (LLR) calculation unit calculates an LLR with respect to the bits based on the Euclidean distance between the reception signal and the two candidates.

Description

Carry out the method and apparatus of soft demapping

The application requires to be submitted to the 10-2013-0027779 korean patent application of Department of Intellectual Property of Korea S and to be submitted to the rights and interests of the 10-2013-0146507 korean patent application of Department of Intellectual Property of Korea S on November 28th, 2013 on March 15th, 2013, and the whole of described application are openly incorporated herein by reference for all objects.

Technical field

Below describe and relate to a kind of communication system based on rotating orthogonal amplitude modulation(PAM) (QAM), more particularly, relate to a kind of method and apparatus of for the communication system based on rotation QAM, the symbol receiving being carried out soft demapping.

Background technology

Wireless communication system and broadcast system can use rotating orthogonal amplitude modulation(PAM) (QAM) transmission technology.For example, digital video broadcasting (DVB) standard (for example, digital video broadcasting-second generation ground (DVB-T2) standard) has specified to send data based on rotation QAM.

In the wireless channel of wireless communication system, can decline.Here, decline refers to the decay of the wireless signal between the transmission period of wireless signal.In order to overcome fading problem, can adopt diversity (diversity) method, deversity scheme comprises space diversity and frequency diversity.The a lot of technology that comprise DVB-T2 standard are adopting signal space diversity method to solve fading problem.As example, rotation qam constellation figure can be used for realizing signal space diversity method.But in the time that transmitter uses rotation QAM, receiver is carried out soft demapping by carrying out complicated processing.

Summary of the invention

The selection of content of the present invention with the concept of introducing reduced form is provided, in embodiment below, further describes described selection.Content of the present invention is not intended to key feature or the essential feature of the theme of Identification Demand protection, is also not intended to be used as helping to determine the scope of claimed theme.

According to illustrative configuration, a kind of method of carrying out soft demapping is provided, described method comprises: the symbol that represents multiple bits from receiving signal acquisition; Constellation point from the planisphere of the each bit for multiple bits is selected at least two candidates; Based on the Euclidean distance receiving between signal and described at least two candidates, calculate the log-likelihood ratio (LLR) for the each bit in multiple bits.

Described method also can comprise: calculate the Euclidean distance receiving between signal and described at least two candidates.

Described method also can comprise: use the LLR for the each bit in multiple bits, detect multiple bits.

Select at least two candidates' step to comprise: being the first logical value, to select at least one candidate at the each bit being included in the multiple bits in the constellation point in planisphere, is the second logical value, to select at least one another candidate at the each bit being included in the multiple bits in the constellation point in planisphere.

Select at least two candidates' step to comprise: corresponding with M-quadrature amplitude modulation (QAM) symbol in response to symbol, comprise from many in the constellation point that the line of individual constellation point comprises, for every line options candidate in many lines, wherein, " M "=2 ⁿand " n " is greater than two integer.

Constellation point in many lines can be corresponding with the logical value of corresponding bits.

Described method also can comprise: by the each index stores in the Euclidean distance between the reception signal of described bit and described at least two candidates and described at least two candidates in look-up table (LUT).

The step of calculating LLR can comprise: use the each index in Euclidean distance and described at least two candidates that are stored between reception signal in LUT, described bit and described at least two candidates, the LLR of the bit in counting period.

The bit at described bit and described interval can be the bit of odd-numbered or the bit of even-numbered.

Described method also can comprise: the planisphere that rotation receives signal is in opposite direction to provide described planisphere according to vertical parallel or horizontal parallel.

According to illustrative configuration, a kind of equipment of carrying out soft demapping is provided, described equipment comprises: pretreatment unit, is configured to represent the symbol of multiple bits from receiving signal acquisition; Candidate's selected cell, the constellation point being configured to from the planisphere of the each bit for multiple bits is selected at least two candidates; Log-likelihood ratio (LLR) computing unit, is configured to the Euclidean distance based on receiving between signal and described at least two candidates, calculates the LLR for the each bit in multiple bits.

Described equipment also can comprise: metrics calculation unit, is configured to calculate the Euclidean distance receiving between signal and described at least two candidates.

Described equipment also can comprise: final full LLR computing unit, be configured to use the LLR for the each bit in multiple bits, and calculate the final full LLR for detection of multiple bits.

Candidate's selected cell can be configured to: being the first logical value, to select at least one candidate at the each bit being included in the multiple bits in the constellation point in planisphere, is the second logical value, to select at least one another candidate at the each bit being included in the multiple bits in the constellation point in planisphere.

Candidate's selected cell can be configured to: corresponding with M-quadrature amplitude modulation qam symbol in response to symbol, comprise from many in the constellation point that the line of individual constellation point comprises, for every line options candidate in many lines, wherein, " M "=2 ⁿand " n " is greater than two integer.

Each index in Euclidean distance and described at least two candidates between the reception signal of bit and described at least two candidates can be stored in look-up table (LUT).LLR computing unit can be configured to use the each index in Euclidean distance and described at least two candidates that are stored between reception signal in LUT, described bit and described at least two candidates, the LLR of the bit in counting period.

The bit at described bit and described interval can be the bit of odd-numbered or the bit of even-numbered.

Described equipment also can comprise: rotary unit, is configured to rotate in opposite direction the planisphere that receives signal to provide described planisphere according to vertical parallel or horizontal parallel.

According to illustrative configuration, a kind of method of carrying out soft demapping is provided, described method comprises: based on receiving signal, in the planisphere of the bit in multiple bits of the symbol for for transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point; Based on the candidate constellation points that receives signal and selection, calculate the log-likelihood ratio (LLR) of described bit.

Corresponding with M-quadrature amplitude modulation (QAM) symbol in response to symbol, the quantity of the candidate constellation points of selection can be wherein, " M "=2 ⁿand " n " is greater than two integer.

Symbol can be with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

The step of calculating the LLR of described bit can comprise: calculate the distance receiving between signal and the candidate constellation points of selection; Based on the distance of calculating, calculate the LLR of described bit.

Many line can comprise: with the corresponding line of the first logical value of described bit and with the corresponding line of the second logical value of described bit.

Corresponding with M-QAM symbol in response to symbol, every line in many lines by individual constellation point forms, wherein, and " M "=2 ⁿand " n " is greater than two integer.

According to another illustrative configuration, a kind of method of carrying out soft demapping is provided, described method comprises: based on receiving signal, in the planisphere of the bit in the multiple bits that comprise for M-quadrature amplitude modulation (QAM) symbol, select individual candidate constellation points, wherein, " M "=2 ⁿand " n " is greater than two integer; Based on the candidate constellation points that receives signal and selection, calculate the log-likelihood ratio (LLR) of described bit.

The step of calculating the LLR of described bit can comprise: calculating reception signal and selection distance between individual candidate constellation points; Based on the distance of individual calculating is calculated the LLR of described bit.

Symbol can be with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

According to illustrative configuration, a kind of equipment of carrying out soft demapping is provided, described equipment comprises: selected cell, is configured to, based on receiving signal, in the planisphere of the bit in the multiple bits that comprise for M-quadrature amplitude modulation (QAM) symbol, select individual candidate constellation points, wherein, " M "=2 ⁿand " n " is greater than two integer; Computing unit, is configured to the candidate constellation points based on receiving signal and selection, calculates the log-likelihood ratio (LLR) of described bit.

According to another illustrative configuration, a kind of equipment of carrying out soft demapping is provided, described equipment comprises: selected cell, be configured to based on receiving signal, in the planisphere of the bit in the multiple bits for comprising for the symbol of transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point; Computing unit, is configured to the candidate constellation points based on receiving signal and selection, calculates the log-likelihood ratio (LLR) of described bit.

It is the bit of even-numbered in response to described bit, the candidate constellation points that computing unit can be configured to the selection that uses described bit is calculated the LLR of the bit of another even-numbered, be the bit of odd-numbered in response to described bit, the candidate constellation points that computing unit can be configured to the selection that uses described bit is calculated the LLR of the bit of another odd-numbered.

Computing unit can be configured to calculate the distance receiving between signal and the candidate constellation points of selection, and distance based on calculating is calculated the LLR of described bit.

Corresponding with M-QAM symbol in response to symbol, wherein, the quantity of the candidate constellation points of selection can be , wherein, " M "=2 ⁿand " n " is greater than two integer.

Symbol can be with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

Described equipment also can comprise: gradient calculation unit, the anglec of rotation and the channel status that are configured to based on planisphere calculate the gradient for the reference line in the planisphere of individual bit.

Selected cell also can be configured to the gradient based on receiving signal and reference line, in the planisphere of the bit in the multiple bits for comprising for the symbol of transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point.

Full LLR computing unit can be configured to use the LLR of multiple bits to calculate the full LLR for detection of multiple bits.

From detailed description, accompanying drawing and claim below, other features and aspect will be clearly.

Brief description of the drawings

From the following description of embodiment by reference to the accompanying drawings, these and/or other aspect will become clear and more easily be understood, wherein:

Figure 1A is the diagram that the example of the rotating orthogonal amplitude modulation(PAM) (QAM) of Q delay has been shown according to the application of embodiment.

Figure 1B and Fig. 1 C are the diagrams illustrating according to the diversity gain producing in the time that use has the rotation qam symbol of Q delay of embodiment.

Fig. 2 A is the diagram illustrating according to the example of the original 16QAM of embodiment, rotation 16QAM and decline 16QAM.

Fig. 2 B is the diagram illustrating according to the example of the planisphere for multiple bits of embodiment.

Fig. 3 A is the diagram that the example of vertical 16QAM planisphere that the log-likelihood ratio for multiple bits (LLR) according to embodiment calculates and horizontal 16QAM planisphere is shown.

Fig. 3 B is the diagram illustrating according to the non-perpendicular 16QAM planisphere of the calculating of the LLR for multiple bits of embodiment or the example of non-horizontal 16QAM planisphere.

Fig. 4 A is the diagram of selecting the candidate's who calculates for LLR example in vertical 16QAM planisphere and horizontal 16QAM planisphere illustrating according to embodiment.

Fig. 4 B is the diagram of selecting the candidate's who calculates for LLR example in non-perpendicular 16QAM planisphere or non-horizontal 16QAM planisphere illustrating according to embodiment.

Fig. 5 is the block diagram illustrating according to the example of the soft demapping device of embodiment.

Fig. 6 is the flow chart illustrating according to the example of the soft de-mapping method of embodiment.

Fig. 7 A to Fig. 7 D is the diagram illustrating according to the example of the method for many lines (line) selection optimal candidate of embodiment.

Fig. 8 A and Fig. 8 B are the diagrams illustrating according to calculating the example of the method for the gradient of reference line in the planisphere of the bit of the even-numbered (even-numbered) in the multiple bits for being included in symbol of embodiment.

Fig. 9 is the block diagram illustrating according to the example of the soft demapping device of embodiment.

Figure 10 is the flow chart illustrating according to the example of the soft de-mapping method of embodiment.

At whole accompanying drawing with in describing in detail, unless otherwise described or regulation, otherwise identical drawing reference numeral will be understood to represent identical element, feature and structure.Accompanying drawing can not to scale (NTS) convergent-divergent, for clear, illustrates and convenient, and relative size, ratio and the description of the element in accompanying drawing can be exaggerated.

Embodiment

Provide following detailed description to help reader to obtain the complete understanding to method described here, equipment and/or system.But various changes, amendment and the equivalent of system described here, equipment and/or method will be obvious for the ordinary skill in the art.In addition,, for more clear and succinct, can omit the description for the known function of those of ordinary skill in the art and structure.

At whole accompanying drawing with in describing in detail, identical label represents identical element.Accompanying drawing can not to scale (NTS) convergent-divergent, for clear, illustrates and convenient, and relative size, ratio and the description of the element in accompanying drawing can be exaggerated.

Can implement feature described here with different forms, and described feature will not be construed as limited to example described here.More relevantly, provide example described here, thereby the disclosure will be thorough and complete, and four corner of the present disclosure is conveyed to those of ordinary skill in the art.

Figure 1A has illustrated according to the application of embodiment the example of the rotating orthogonal amplitude modulation(PAM) (QAM) of Q delay.In general, 4-QAM comprises the constellation point 111,112,113 and 114 being arranged in the foursquare constellation Figure 110 with identical Q channel spacing and identical I channel spacing.Each and multiple bit in constellation point 111,112,113 and 114 is corresponding.For example, constellation point 111 is corresponding with " 00 ", and constellation point 112 is corresponding with " 01 ", and constellation point 113 is corresponding with " 11 ", and constellation point 114 is corresponding with " 10 ".

As " n " individual bit { b ₀, b ₁, b ₂..., b _n-1(" n " is greater than two integer) be assumed to be and be mapped to M-QAM symbol (M=2 ⁿ) time, M-QAM symbol is by S=S _i+ jS _qrepresent, wherein, " s " represents M-QAM symbol, " S _i" represent the signal of I channel, " S _q" represent the signal of Q channel, " j " represents the operator of imaginary part.In the following description, unless otherwise stated, otherwise " qam symbol " comprises the general M-QAM symbol of for example qam symbol, 16-QAM symbol, 32-QAM symbol, 64-QAM symbol or other type.

Rotation 4-QAM uses the rotation constellation Figure 119 obtaining by constellation Figure 110 being rotated to predetermined angular " α ".Rotation constellation Figure 119 comprises according to the constellation point 115,116,117 and 118 of predetermined angular " α " rotation.Each and multiple bit in rotation constellation point 115,116,117 and 118 is corresponding.For example, constellation point 115 is corresponding with " 00 ", and constellation point 116 is corresponding with " 01 ", and constellation point 117 is corresponding with " 11 ", and constellation point 118 is corresponding with " 10 ".When produce rotation qam symbol with anglec of rotation α, by equation 1 definition rotation qam symbol.According to embodiment, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere that is included in the bit in the symbol that Q postpones.

[equation 1]

$\begin{array}{c}\tilde{s}=e^{\mathrm{j\α}}s={\tilde{s}}_I+j{\tilde{s}}_Q\\ =(s_I\cos \mathrm{\α}-s_Q\sin \mathrm{\α})+j(s_I\sin \mathrm{\α}+s_Q\cos \mathrm{\α})\end{array}$

In equation 1, represent rotation qam symbol, e ^{j α}represent rotation operator, represent the rotating signal by I channel signal rotation anglec of rotation α is obtained, represent the rotating signal by Q channel signal rotation anglec of rotation α is obtained.

Processor receives signal from the rotation qam symbol sending from transmitter.Processor is carried out OFDM (OFDM) solution and is transferred from receiving the multiple bits of input.For example, processor is from collecting real part with the corresponding subcarrier of I channel, and from collecting imaginary part with the corresponding subcarrier of Q channel.Use real part and the imaginary part of collecting from two corresponding subcarriers at processor to represent to receive signal by equation 2.

[equation 2]

$r=r_I+j{r}_Q=h_I{\tilde{s}}_I+j{h}_Q{\tilde{s}}_Q+w$

In equation 2, " r " represents to receive signal, " r _i" represent the reception signal of I channel, " r _q" represent the reception signal of Q channel, h _ithe channel information of expression and the corresponding subcarrier of I channel, h _qrepresent and the channel information of the corresponding subcarrier of Q channel, " w " represents noise.Each channel information represents the amplitude of channel.

In general, be included in the I channel signal of the real part in qam symbol and the Q channel signal of imaginary part via the identical carrier wave transmission of the phase difference with 90 degree.For example, I component and the Q component of rotation qam symbol are indicated as frame 121.The every row that are included in frame 121 are indicated single carrier wave.In the time not applying Q delay, in OFDM, will rotate qam symbol " s ₁" real part " I ₁" 123 and imaginary part " Q ₁" 124 be mapped to identical subcarrier.

Q is postponed to be incorporated into rotation QAM and obtain extra diversity.Q postpones to represent such method: will be included in the I channel signal of the real part in identical qam symbol and the Q channel signal of imaginary part is mapped to different carrier waves.In the time that Q is postponed to be incorporated into rotation QAM, can carry out component axle and interweave.In the time that Q is postponed to be incorporated into rotation qam symbol, in OFDM, will rotate real part and the imaginary part mapping of qam symbol and send to multiple different subcarriers.For example, I component and the Q component of having introduced the rotation qam symbol that Q postpones are indicated as frame 122.In the time introducing Q delay, will rotate respectively qam symbol " s ₁" real part " I ₁" 125 and imaginary part " Q ₁" 126 be mapped to multiple different subcarriers.Circulation Q postpones to be introduced in forward error correction (FEC) module unit.As shown in frame 122, the last Q channel signal " Q in frame _n" and an I channel signal " I ₁" postpone to be mapped to identical carrier wave by circulation Q.

When introducing Q and postpone to carry out component axle and interweave, processor recovers the signal of decline unit from the signal of different units.Decline unit is the scheduled unit that does not receive signal due to decline.In an example, individual unit is corresponding with single OFDM subcarrier.When introducing Q and postpone to carry out component axle and interweave, because at " I ₁" 125 unit that are mapped to and " Q ₁" there is decline in 126 unit that are mapped to probability relatively lower than at " I ₁" 125 unit that are mapped to and " Q ₁" there is the probability of decline in 126 unit that are mapped to, so processor does not receive " I ₁" 125 and " Q ₁" 126 both probability are low significantly.Therefore,, when do not receive one of the real part signal of predetermined rotation qam symbol and imaginary part signal due to decline, processor uses another in described two signals to recover not receive signal.For example,, when because decline 127 does not receive " Q ₁" 126 o'clock, receiver uses " I ₁" 125 recovery " Q ₁" 126.

Figure 1B and Fig. 1 C illustrate the diagram of the example of the situation of the diversity gain producing in the time that use has the rotation qam symbol that Q postpones compared with using the situation of general qam symbol.With reference to Figure 1B, in the time using general qam symbol, transmitter sends and is included in the corresponding symbol of one of constellation point 131,132,133 and 134 in constellation Figure 130.For example, transmitter sends with the corresponding symbol of constellation point 131 and sends two bits (for example, " 10 ").Selectively, transmitter sends with the corresponding symbol of constellation point 132 and sends two bits (for example, " 11 ").Because distinguish by Q channel with the corresponding symbol of constellation point 131 with the corresponding symbol of constellation point 132, so in response to declining in Q channel, processor be difficult to determine from transmitter send symbol be with the corresponding symbol of constellation point 131 or with the corresponding symbol of constellation point 132.

With reference to Fig. 1 C, in the time using rotation qam symbol, transmitter sends and is included in the corresponding symbol of one of constellation point 151,152,153 and 154 in constellation Figure 150.As use general qam symbol situation substitute, no matter whether decline in Q channel, processor all use rotation qam symbol recover from transmitter send symbol.For example, whether no matter decline in Q channel, in the time sending with constellation point 151 corresponding symbol, processor all receives and the corresponding signal that receives of constellation point 161.In response to receiving accordingly signal with constellation point 161, processor is defined as the symbol sending from transmitter and the corresponding symbol of constellation point 151.Send and the corresponding symbol of constellation point 152 in response to transmitter, whether no matter decline in Q channel, receiver all receives and the corresponding signal that receives of constellation point 162.In the time receiving with the corresponding reception signal of constellation point 162, processor is defined as the symbol sending from transmitter and the corresponding symbol of constellation point 152.Like this, compared with using the situation of general qam symbol, in the time using rotation qam symbol, produce diversity gain.

Fig. 2 A illustrates according to the example of the original 16QAM of embodiment, rotation 16QAM and decline 16QAM.Below, describe with being easy to for simplicity, use the example of 16QAM; But the various details that belong to the description of 16QAM can be applicable to for example 64QAM or 256QAM and non-square 32QAM or 128QAM.

With reference to Fig. 2 A, original 16QAM planisphere 210 comprises 16 constellation point.The bit-order of original 16QAM planisphere 210 is assumed to be b ₀-b ₁-b ₂-b ₃.The bit-order of original 16QAM planisphere 210 can be different order.In the multiple bits that are included in symbol, bit (for example, the b of even-numbered ₁or b ₃) index (for example, " 1 " or " 3 ") and Odd Phases should.Bit (for example, the b of odd-numbered ₀or b ₂) index (for example, " 0 " or " 2 ") corresponding with even number.Be included in each information representing with four bits that comprises of 16 constellation point in original 16QAM planisphere 210.For example, in 16 constellation point each with corresponding to " 1111 " from " 0000 ".Each corresponding based on being arranged in position in original 16QAM planisphere 210 and from " 0000 " to " 1111 " one in 16 constellation point.As an example, based on Gray code, with contiguous constellation point on original 16QAM planisphere 210 accordingly the bit from " 0000 " to " 1111 " be assigned to respectively 16 constellation point, to allow single potential difference (single digit difference) in four bit values.Bit information shown in Fig. 2 A is only exemplary, and can be modified to various forms.

In the time that original 16QAM planisphere 210 is rotated to the anglec of rotation " α ", obtain rotation 16QAM planisphere 220.Transmitter uses 16 constellation point that are included in 16QAM planisphere 220 to send data.For example, transmitter sends with the corresponding symbol of constellation point 221 to send " 0101 ".Constellation point 221 with by from original 16QAM planisphere 210 will with " 0101 " accordingly constellation point 211 to rotate the rotation constellation point that the anglec of rotation " α " obtains corresponding.

In the time rotating 16QAM planisphere 220 decline in I channel and/or Q channel, obtain the rotation 16QAM planisphere 230 of decline.In the time that the rotation 16QAM of Q delay has been introduced in use, the degree of I channel is different from the fading severity of Q channel.In an example, the rotation 16QAM planisphere 230 of decline is the size " h of I channel _i" be the size " h of " 0.9 " and Q channel _q" be the planisphere of the situation of " 0.4 ".In the time that the constellation point of the rotation 16QAM planisphere 230 of decline is compared with the constellation point of rotation 16QAM planisphere 220, because the size " h of the I channel of the rotation 16QAM planisphere 230 of decline _i" be " 0.9 ", so described constellation point is dwindled to arrange described constellation point with the corresponding x direction of principal axis of I channel according to the degree of " 0.9 " by edge.In the time that the constellation point of the rotation 16QAM planisphere 230 of decline is compared with the constellation point of rotation 16QAM planisphere 220, because the size " h of the Q channel of the rotation 16QAM planisphere 230 of decline _q" be " 0.4 ", so described constellation point is dwindled to arrange described constellation point with the corresponding y direction of principal axis of Q channel according to the degree of " 0.4 " by edge.Therefore, the rotation 16QAM planisphere 230 of decline is no longer provided with regular shape.

Rotation 16QAM planisphere 220 can be the planisphere using in transmitter, and the rotation 16QAM planisphere 230 of decline can be the planisphere using in processor or receiver.For example, in the time that the rotation 16QAM symbol producing based on rotation 16QAM planisphere 220 experiences decline in the time being sent out, the rotation 16QAM planisphere 230 of processor based on decline recovers to receive signal.

Can be used for from receiving Signal estimation bit information for the equipment (following, to be also called as soft demapping device) of carrying out soft demapping.Soft demapping device produces the information for the multiple bits that send from transmitter are decoded from receiving signal.For example, soft demapping device uses and receives the log-likelihood ratio (LLR) of calculated signals for multiple bits.Be used for multiple bits to decode for the LLR of multiple bits.

Fig. 2 B illustrates according to the method for the calculating LLR of embodiment and for the example of the planisphere of multiple bits.

With reference to Fig. 2 B, for individual bit b ₀planisphere 240, for individual bit b ₁planisphere 250, for individual bit b ₂planisphere 260 and for individual bit b ₃planisphere 270 comprise the constellation point that is arranged in the position that the constellation point that comprises with the rotation 16QAM planisphere of the decline of Fig. 2 A is identical.Be included in for individual bit b ₀planisphere 240 in constellation point, for individual bit b ₁planisphere 250, for individual bit b ₂planisphere 260 and for individual bit b ₃the logical value of planisphere 270 based on corresponding bits be partitioned.In Fig. 2 B, solid dot ● represent to pass through b _i=1 constellation point defining, hollow out point zero represents from b _i=0 constellation point deriving.

For example, be included in for individual bit b ₀planisphere 240 in constellation point be partitioned individual bit b ₀logical value be subset 241 and the individual bit b of " 1 " ₀logical value be the subset 242 of " 0 ".For the each individual bit b in constellation point ₀logical value with in the original 16QAM planisphere 210 of Fig. 2 A instruction the first bit b ₀logical value identical.Be included in for individual bit b ₁planisphere 250 in constellation point be partitioned individual bit b ₁logical value be subset 251 and the individual bit b of " 1 " ₁logical value be the subset 252 of " 0 ".Individual bit b in each in constellation point ₁logical value with in the original 16QAM planisphere 210 of Fig. 2 A instruction the second bit b ₁logical value identical.

Be included in for individual bit b ₂planisphere 260 in constellation point be partitioned individual bit b ₂logical value be subset 262 and the individual bit b of " 1 " ₂logical value be the subset 263 of " 0 ".Individual bit b in each in constellation point ₂logical value with in the original 16QAM planisphere 210 of Fig. 2 A instruction the 3rd bit b ₂logical value identical.Be included in for individual bit b ₃planisphere 270 in constellation point be partitioned individual bit b ₃logical value be subset 271 and the individual bit b of " 1 " ₃logical value be the subset 272 and 273 of " 0 ".Individual bit b in each in constellation point ₃logical value with in the original 16QAM planisphere 210 of Fig. 2 A instruction the 4th bit b ₃logical value identical.Partition information shown in Fig. 2 B is only exemplary, the change for the subregion of the planisphere 240,250,260 and 270 of individual bit in can the bit information based on shown in Fig. 2 A and changing.

When using in the situation of general QAM relatively, comprise large computation complexity for receiver by the situation of rotating QAM and carry out soft demapping.In the situation that using general QAM, because I component and Q component are independent of one another, so can calculate LLR by one dimension (1D).On the contrary, in the time using rotation QAM, due to the correlation between I component and Q component, can need two dimension (2D) to calculate for all constellation point that are included in planisphere.

In the time using rotation QAM, for the each bit b in the multiple bits that are included in symbol _icalculate the LLR that receives signal " r ", to carry out soft demapping.Define for individual bit b by equation 3 _ilLR.

[equation 3]

In equation 3, LLR(b _i) represent for the LLR of " i " the individual individual bit in the multiple bits that are included in symbol.σ ²represent the amplitude of noise, represent and b _i=1 relevant subset, represent and b _i=0 relevant subset.

Approximate based on MAX-log, LLR can represent by equation 4.

[equation 4]

With reference to equation 4, for individual bit b _ilLR calculate depend on that two set that are different in essence are (for example,, with b _i=1 relevant set and and b _i=0 relevant set) in constellation point and receive the Euclidean distance between signal " r ".With b _idistance in=0 relevant set receive the nearest constellation point of signal " r " and with b _idistance in=1 relevant set receives the nearest constellation point of signal " r " need to be determined, to calculate for individual bit b based on equation 4 _ilLR.Calculating the distance receiving between signal " r " and all constellation point is to detect the direct method that sends bit for LLR calculating.But such full-search algorithm can be complicated and huge for realization.

Soft demapping device is configured to: in the case of calculating the distance receiving between signal " r " and all constellation point, calculate the LLR that has and calculate the essentially identical accuracy of accuracy of the situation that receives the distance between signal " r " and all constellation point.Soft demapping device can be by after the single candidate constellation points of many line options for being formed by the multiple constellation point in planisphere, calculates the candidate constellation points of selecting and receives the distance between signal, calculates LLR.After a while description is belonged to the further describing of method of selecting candidate constellation points.

Fig. 3 A illustrates according to the vertical 16QAM planisphere of the calculating of the LLR for multiple bits of embodiment and the example of horizontal 16QAM planisphere.With reference to Fig. 3 A, soft demapping device is the planisphere 240,250,260 and 270 of rotation diagram 2B in opposite direction, be arranged with the form of horizontal parallel shape or vertical parallel shape with the constellation point that allows to be included in planisphere 240,250,260 and 270, to realize the selection of candidate constellation points.For example, soft demapping device obtains for individual bit b by the planisphere 240 of rotation diagram 2B in opposite direction ₀planisphere 310.Soft demapping device obtains for individual bit b by the planisphere 250 of rotation diagram 2B in opposite direction ₁planisphere 320.Soft demapping device also obtains for individual bit b by the planisphere 260 of rotation diagram 2B in opposite direction ₂planisphere 330.In addition, soft demapping device obtains for individual bit b by the planisphere 270 of rotation diagram 2B in opposite direction ₃planisphere 340.

Be included in for individual bit b ₀planisphere 310 in constellation point, for individual bit b ₁planisphere 320, for individual bit b ₂planisphere 330, for individual bit b ₃planisphere 340 in the logical value of constellation point based on corresponding bits be partitioned.In Fig. 3 A, ● represent to pass through b _i=1 constellation point defining, zero represents from b _i=0 constellation point deriving.Be included in for individual bit b ₀planisphere 310 in constellation point be partitioned individual bit b ₀logical value be subset 311 and the individual bit b of " 1 " ₀logical value be the subset 312 of " 0 ".For individual bit b ₁planisphere 320 in constellation point be partitioned individual bit b ₁logical value be subset 321 and the individual bit b of " 1 " ₁logical value be the subset 322 of " 0 ".For individual bit b ₂planisphere 330 in constellation point be partitioned individual bit b ₂logical value be subset 331 and the individual bit b of " 1 " ₂logical value be the subset 332 and 333 of " 0 ".For individual bit b ₃planisphere 340 in constellation point be partitioned individual bit b ₃logical value be subset 341 and the individual bit b of " 1 " ₃logical value be the subset 342 and 343 of " 0 ".Partition information shown in Fig. 3 A is only exemplary.Can be partitioned the different subset all with the logical value different from the logical value shown in Fig. 3 A for the constellation point in each planisphere of individual bit.In addition the change in the bit information based on shown in Fig. 2 A and changing aspect rotation for the subregion of planisphere 310,320,330 and 340 of the rotation in opposite direction of individual bit.

But subregion may needn't require the planisphere receiving to rotate in opposite direction.In addition the example providing at this, also can be applicable to planisphere and the erose planisphere of vertical parallel shape or horizontal parallel shape.

By rotating in opposite direction, for individual bit b ₀planisphere 310 and for individual bit b ₂planisphere 330 there is the constellation point of vertical parallel.The constellation point in identical row about the constellation point of vertical parallel shape belongs to identical subset.For example,, for individual bit b ₀planisphere 310 in, all belong to subset 311 with the constellation point on the corresponding Article 1 line 313 of first row, all belong to subset 311 with the constellation point on the corresponding Article 2 line 314 of secondary series.All belong to subset 312 with the constellation point on the corresponding Article 3 line 315 of the 3rd row, all belong to subset 312 with the constellation point on the corresponding Article 4 line 316 of the 4th row.

For individual bit b ₁planisphere 320 and for individual bit b ₃planisphere 340 there is the constellation point of horizontal parallel shape.Belong to identical subset about the constellation point in the identical row of the constellation point of horizontal parallel.For example,, for individual bit b ₃planisphere 340 in, belong to subset 342 with the constellation point on the corresponding Article 5 line 344 of the first row.In addition, with the corresponding Article 6 line 345 of the second row on constellation point belong to subset 341.Belong to subset 341 with the constellation point on the corresponding Article 7 line 346 of the third line.In addition, with the corresponding Article 8 line 347 of fourth line on constellation point all belong to subset 343.Line shown in Fig. 3 B is only exemplary, the change for the line of the planisphere 350,360,370 and 380 of individual bit in can the bit information based on shown in Fig. 2 A and revising.

Due to so above-mentioned configuration, can simplify soft demapping algorithm.Soft demapping device is for the bit of odd-numbered, from planisphere 310 and 330 for every column selection apart from receiving signal nearest single candidate, and receive signal nearest single candidate from planisphere 320 and 340 for every row chosen distance for the bit of even-numbered.By way of example, in vertical parallel shape, determine the optimal candidate in single row, for example, have to the constellation point of the minimum range of reception signal.

Soft demapping device does not need to calculate the distance receiving between signal and constellation point to select candidate.Reception signal is vertically projected to every row or every row by soft demapping device, and the nearest constellation point of point that distance reception signal is projected to is elected to be the candidate for corresponding line or respective column.For example, by vertically being projected to respective column, reception signal determines the optimal candidate of single row.Soft demapping device is vertically projected to distance reception signal the nearest constellation point of point of respective column, is defined as the optimal candidate for respective column.Soft demapping device is determined optimal candidate for multiple row, obtains by collecting optimal candidate for every row the candidate collection that quantity reduces.

After selecting candidate, the candidate that soft demapping device is selected by calculating and the distance receiving between signal are calculated LLR.For example, can calculate the Euclidean distance receiving between signal " r " and optimal candidate.When using M-QAM symbol (wherein, " M "=2 ⁿ, " n " is greater than two integer) time, by calculating individual constellation point and receive the distance between signal instead of calculate the distance between " M " individual constellation point and reception signal altogether, calculates LLR.

Have and reduce by usage quantity the constellation point of the minimum range that individual candidate collection obtains is with to have constellation point by the use minimum range that the set of " M " individual point obtains altogether identical.Therefore, soft demapping device is configured to calculate accurate LLR in the situation that lacking corner (corner), and does not need to calculate all constellation point and receive the 2D distance between signal.The further description of the soft demapping algorithm that belongs to simplification is discussed with reference to Fig. 4 A and Fig. 4 B after a while.

Fig. 3 B illustrates according to the non-perpendicular 16QAM planisphere of the calculating of the LLR for multiple bits of embodiment or the example of non-horizontal 16QAM planisphere.With reference to Fig. 3 B, as mentioned above, should notice that vertical parallel shape or horizontal parallel shape are illustrative objects.For individual bit b ₀planisphere 350 and for individual bit b ₂planisphere 370 have from vertical parallel shape in the counterclockwise direction according to the constellation point of the form of predetermined angular rotation.About always belonging to identical subset according to the constellation point on identical line of the constellation point of the form of predetermined angular rotation in the counterclockwise direction from vertical parallel shape.For individual bit b ₁planisphere 360 and for individual bit b ₃planisphere 380 have from horizontal parallel shape in the counterclockwise direction according to the constellation point of the form of predetermined angular rotation.About always belonging to identical subset according to the constellation point on identical line of the constellation point of the form of definite angle rotation in the counterclockwise direction from horizontal parallel shape.As about Fig. 3 A as described in front, due to such characteristic, soft demapping algorithm can be simplified.

Fig. 4 A illustrates the example of selecting the candidate who calculates for LLR in vertical 16QAM planisphere and horizontal 16QAM planisphere according to embodiment.With reference to Fig. 4 A, each in four row in planisphere 410 and planisphere 420 comprises four constellation point, and each in four row in planisphere 430 and 440 comprises four constellation point.

In illustrated examples, planisphere 410 has the constellation point of vertical parallel shape.In this example, can easily determine single optimal candidate for multiple row.In the case of not calculating four constellation point in single row and receiving four distances between signal, determine optimal candidate.Reception signal 415 is vertically projected to every row by soft demapping device, and the nearest constellation point of point that distance reception signal 415 is vertically projected to is elected to be the optimal candidate for respective column.Be pre " d because be included in interval between the constellation point in the every row in planisphere 410 _v" 416, so soft demapping device is easily determined the constellation point nearest apart from subpoint.For example, receiving the point 412 that signal 415 is projected to row 411 is disposed between constellation point (2,1) and constellation point (3,1).Because constellation point (2,1) is " d with the interval between constellation point (3,1) _v" 416, and subpoint 412 arrives constellation point (3,1) at distance " d _v/ 2 " within, so be confirmed as constellation point (3,1) for the optimal candidate of row 411.In a similar fashion, the optimal candidate in planisphere 410 is confirmed as (3,1), (3,2), (2,3) and (2,4).The single optimal candidate of soft demapping device based on for every row calculated the LLR for corresponding bits.

Soft demapping device is implemented as simplifies whole LLR computing.With reference to the equation 4 providing above, b ₀logical value be that distance in the subset of " 1 " receives nearest constellation point and the b of signal 415 ₀logical value need to be determined for the distance in the subset of " 0 " receives the nearest constellation point of signal 415, to calculate for individual bit b ₀lLR.Soft demapping device by calculate for four optimal candidate of every column selection with receive distance between signal and calculate the LLR that has and calculate the accuracy that the accuracy of the distance between all constellation point and reception signal is identical.In planisphere 410, b ₀logical value to be that distance in the subset of " 1 " receives the nearest constellation point of signal 415 corresponding with constellation point (3,2), b ₀logical value to be that distance in the subset of " 0 " receives the nearest constellation point of signal 415 corresponding with constellation point (2,3).

In this example, the candidate collection that usage quantity reduces jointly.With reference to planisphere 410 and the planisphere 420 of Fig. 4 A, b ₀and b ₂candidate mutually the same.For example, constellation point (3,1), (3,2), (2,3) and (2,4) are b ₀and b ₂candidate.Difference is the each different subset belonging to for multiple different bits in candidate.For example, (3,1) belong in planisphere 410 with " b ₀=1 " relevant subset, (3,1) belong in planisphere 420 with " b ₂=0 " relevant subset.Distance for the precomputation between index and candidate and the reception signal of the candidate in the candidate collection that is included in quantity minimizing of predetermined bit also can be used as look-up table (LUT) for other bits.For example, calculate for b ₀the optimal candidate constellation point (3,1), (3,2), (2,3) and (2,4) of every row in each and receive the distance between signal.By the index record of constellation point (3,1), (3,2), (2,3) and (2,4) in LUT.Distance between constellation point (3,1), (3,2), (2,3) and (2,4) and reception signal is recorded in LUT.LUT can be reused for and calculate for bit b ₂the processing of LLR.In planisphere 420, b ₂logical value to be that distance in the subset of " 1 " receives the nearest constellation point of signal corresponding with constellation point (3,2).In addition b, ₂logical value to be that distance in the subset of " 0 " receives the nearest constellation point of signal corresponding with constellation point (3,1).

In a similar fashion, have in the example of constellation point of horizontal parallel shape at planisphere 430, can easily determine single optimal candidate for multiple row.According to example, in the case of not calculating four constellation point in single row and receiving four distances between signal, determine optimal candidate.Reception signal 435 is vertically projected to each row by soft demapping device, and the nearest constellation point of point that distance reception signal 435 is vertically projected to is elected the optimal candidate of corresponding line as.Because the interval between the constellation point in the every row in planisphere 430 is pre " d _n" 436, so soft demapping device is easily determined the constellation point nearest apart from subpoint.In a similar fashion, the optimal candidate in planisphere 430 is confirmed as (1,3), (2,3), (3,2) and (4,2).The single optimal candidate of soft demapping device based on for every row calculated the LLR for corresponding bits.

By this processing, can simplify whole LLR computing.With reference to the equation 4 providing, determine b above ₁logical value be that distance in the subset of " 1 " receives nearest constellation point and the b of signal 435 ₁logical value be that distance in the subset of " 0 " receives the nearest constellation point of signal, to calculate for individual bit b ₁lLR.Soft demapping device passes through to calculate four optimal candidate selecting for every row and receives the distance between signal, calculates the LLR with the accuracy identical with the accuracy of calculating the distance between all constellation point and reception signal.In planisphere 430, b ₁logical value to be that distance in the subset of " 1 " receives the nearest constellation point of signal 435 corresponding with constellation point (3,2), b ₁logical value to be that distance in the subset of " 0 " receives the nearest constellation point of signal 435 corresponding with constellation point (2,3).

In this example, the candidate collection that usage quantity reduces jointly.With reference to planisphere 430 and the planisphere 440 of Fig. 4 A, for b ₁and b ₃candidate mutually the same.For example, constellation point (1,3), (2,3), (3,2) and (4,2) are for b ₁and b ₃candidate.Difference is the each different subset belonging to for multiple different bits in candidate.For example, (4,2) belong in planisphere 430 with " b ₁=1 " relevant subset, (4,2) belong in planisphere 440 with " b ₃=0 " relevant subset.The distance of the precomputation between index and candidate and the reception signal of the candidate in the candidate collection reducing for the quantity of predetermined bit also can be used as LUT for other bits.For example, calculate b ₁the optimal candidate constellation point (1,3), (2,3), (3,2) and (4,2) of every row in each and receive the distance between signal.By the index record of constellation point (1,3), (2,3), (3,2) and (4,2) in LUT.Distance between constellation point (1,3), (2,3), (3,2) and (4,2) and reception signal is recorded in LUT.LUT can be reused for and calculate for b ₃the processing of LLR.In planisphere 440, b ₃logical value to be that distance in the subset of " 1 " receives the nearest constellation point of signal corresponding with constellation point (3,2), b ₃logical value to be that distance in the subset of " 0 " receives the nearest constellation point of signal corresponding with constellation point (4,2).

Fig. 4 B illustrates the example of selecting the candidate who calculates for LLR in non-perpendicular 16QAM planisphere or non-horizontal 16QAM planisphere according to embodiment.

With reference to Fig. 4 B, in the example providing at this, vertical parallel shape or horizontal parallel shape are optional.Easily determine the single optimal candidate for many lines.For example, reception signal is vertically projected to many lines, determine the single optimal candidate for corresponding line.

Single optimal candidate based on for many lines is calculated the LLR for corresponding bit, instead of calculates for all distances between constellation point and the reception signal of many lines the distance based on calculating and calculate LLR.Therefore, can simplify whole LLR calculates.

With reference to planisphere 450, calculate bit (for example, the b of odd-numbered ₀and b ₂) the optimal candidate constellation point (3,1), (3,2), (2,3) and (2,4) of many lines in each, and receive the distance between signal.By the index record of constellation point (3,1), (3,2), (2,3) and (2,4) in LUT.Distance between constellation point (3,1), (3,2), (2,3) and (2,4) and reception signal is recorded in LUT.LUT can be reused for the processing of calculating LLR for other bits.

With reference to planisphere 460, calculate bit (for example, the b of even-numbered ₁and b ₃) the optimal candidate constellation point (1,3), (2,3), (3,2) and (4,2) of many lines in each, and receive the distance between signal.By the index record of constellation point (1,3), (2,3), (3,2) and (4,2) in LUT.Distance between constellation point (1,3), (2,3), (3,2) and (4,2) and reception signal is recorded in LUT.LUT can be reused for the processing of calculating LLR for other bits.

In general, soft demapping device comprises selected cell and computing unit.Selected cell is based on receiving in the planisphere of the individual bit of signal in multiple bits of the symbol for for transfer of data, for the each selection candidate constellation points in many lines that formed by multiple constellation point.The candidate constellation points of computing unit based on receiving signal and selection calculated the LLR for individual bit.

Fig. 5 illustrates according to the example of the soft demapping device of embodiment.The selected cell of soft demapping device comprises candidate's selected cell 531, and computing unit comprises metrics calculation unit 532 and LLR computing unit 533.Soft demapping device also comprises pretreatment unit 510, final LLR computing unit 550 and post-processing unit 560.

Pretreatment unit 510 receives the reception signal from the symbol of the multiple bits of expression from transmitter.As used herein, symbol can represent M-QAM symbol, wherein, and " M "=2 ⁿand " n " is greater than two integer.According to example, soft demapping device also comprises rotary unit 520.Rotary unit 520 rotates the planisphere that receives signal in opposite direction, to allow planisphere to have vertical parallel shape or horizontal parallel shape.In the time rotating in opposite direction the planisphere that receives signal, as shown in Figure 4 A, the planisphere of reverse rotation has vertical parallel shape or horizontal parallel shape.But with reference to Fig. 4 B, example can require planisphere to have vertical parallel shape or horizontal parallel shape.Below, for being easy to describe, will use the example of such rotary unit 520, described rotary unit rotates the planisphere that receives signal in opposite direction so that planisphere has vertical parallel shape or horizontal parallel shape.

In the time providing planisphere with the form of vertical parallel shape or horizontal parallel shape, processed to received signal individually by the planisphere of the each planisphere for vertical parallel shape in construction unit 530 and piece 540 and horizontal parallel shape.For example, piece 530 calculate for the LLR of the corresponding multiple bits of vertical constellation shapes.Piece 540 calculate for the LLR of the corresponding multiple bits of horizontal constellation shapes.

Candidate's selected cell 531 and 541 is selected at least two candidates from the constellation point the planisphere of rotation in opposite direction for multiple bits.For example, in the planisphere 410 of Fig. 4 A, a single point is selected as for comprising the candidate of every row of individual constellation point.In a similar fashion, in the planisphere 430 of Fig. 4 A, a single point is selected as for comprising the candidate of every row of individual constellation point.Candidate's selected cell 531 and 541 is selected at least one candidate for the each and corresponding situation of the first logical value in the multiple bits in the constellation point in planisphere.Candidate's selected cell 531 and 541 is also selected at least one another candidate for the each and corresponding situation of the second logical value in the multiple bits in the constellation point in planisphere.For example, in the planisphere 410 of Fig. 4 A, constellation point (2,3) and (2,4) are selected as for b ₀bit is the candidate of the situation of " 0 ", and constellation point (3,1) and (3,2) are selected as for b ₀bit is the candidate of the situation of " 1 ".

Metrics calculation unit 532 and 542 is calculated the Euclidean distance receiving between signal and at least two candidates.For example, in the planisphere 420 of Fig. 4 A, calculate the Euclidean distance of constellation point (3,1), (3,2), (2,3) and (2,4), for b ₀lLR calculate.The Euclidean distance that will arrive constellation point (3,1), (3,2), (2,3) and (2,4) is stored in LUT, for b ₂the LLR of bit calculates.Storage constellation point (3,1), (3,2), (2,3) and the index of (2,4) together with to calculate constellation point (3,1), (3,2), the Euclidean distance of (2,3) and (2,4).In a similar fashion, in the planisphere 430 of Fig. 4 A, calculate the Euclidean distance of constellation point (1,3), (2,3), (3,2) and (4,2), for b ₁lLR calculate.The Euclidean distance of the constellation point (1,3) to calculating, (2,3), (3,2) and (4,2) is stored in LUT, for b ₂lLR calculate.Storage constellation point (1,3), (2,3), (3,2) and the index of (4,2) together with to calculate constellation point (1,3), (2,3), the Euclidean distance of (3,2) and (4,2).

The Euclidean distance of LLR computing unit 533 and 543 based on receiving between signal and at least two candidates calculates the LLR for multiple bits.For example, LLR computing unit 533 uses the Euclidean distance to constellation point (3,1), (3,2), (2,3) and (2,4) calculating in metrics calculation unit 532 to calculate b ₀lLR.LLR computing unit 533 uses for b ₀for the shortest constellation point of the Euclidean distance in the constellation point of the situation of " 0 " and for b ₀for the shortest constellation point of Euclidean distance in the constellation point of the situation of " 1 ", calculate b ₀lLR.Euclidean distance to constellation point (2,3) is used for b ₀for the situation of " 0 ", be used for b to the Euclidean distance of constellation point (3,2) ₀for the situation of " 1 ".

In another example, LLR computing unit 543 uses the Euclidean distance to constellation point (1,3), (2,3), (3,2) and (4,2) of calculating in metrics calculation unit 542 to calculate b ₁lLR.LLR computing unit 543 uses for b ₁for the shortest constellation point of the Euclidean distance in the constellation point of the situation of " 0 " and for b ₁for the shortest constellation point of Euclidean distance in the constellation point of the situation of " 1 ", calculate b ₁lLR.Euclidean distance to constellation point (2,3) is used for b ₁for the situation of " 0 ", be used for b to the Euclidean distance of constellation point (3,2) ₁for the situation of " 1 ".

For predetermined bit, by receive Euclidean distance between signal and at least two candidates and for the each index stores in described at least two candidates in LUT.LLR computing unit 533 and 543 uses for the each index in Euclidean distance and described at least two candidates between reception signal and at least two candidates of predetermined bit, calculates the LLR for another bit.In an example, Euclidean distance and index that LLR computing unit 533 use are stored in LUT calculate b ₂lLR.LLR computing unit 533 uses for b ₂for the shortest constellation point of the Euclidean distance in the constellation point of the situation of " 0 " and for b ₂for the shortest constellation point of Euclidean distance in the constellation point of the situation of " 1 ", calculate b ₂the LLR of bit.Euclidean distance to constellation point (3,1) is used for b ₂for the situation of " 0 ", be used for b to the Euclidean distance of constellation point (3,2) ₂for the situation of " 1 ".

As another example, Euclidean distance and index that LLR computing unit 543 use are stored in LUT calculate b ₃lLR.LLR computing unit 543 uses for b ₃for the shortest constellation point of the Euclidean distance in the constellation point of the situation of " 0 " and for b ₃for the shortest constellation point of Euclidean distance in the constellation point of the situation of " 1 ", calculate b ₃lLR.Euclidean distance to constellation point (4,2) is used for b ₃for the situation of " 0 ", be used for b to the Euclidean distance of constellation point (3,2) ₃for the situation of " 1 ".

Final LLR computing unit 550 combines the LLR of multiple bits to detect multiple bits.Will be for { b ₀, b ₂..., b _n-2and { b ₁, b ₃..., b _n-1lLR combination.Post-processing unit 560 uses the LLR of combination to decode to multiple bits.The description of Fig. 1 to Fig. 4 B providing above can be applicable to the each module shown in Fig. 5, for simplicity, will omit the description repeating.

Fig. 6 illustrates according to the example of the soft de-mapping method of embodiment.

650, the distance based between candidate constellation points and reception signal is calculated for the each LLR in multiple bits.

In operation 610, soft de-mapping method carries out preliminary treatment, wherein, receives the reception signal from transmitter from the symbol that represents multiple bits.In operation 621, each in multiple bits of soft de-mapping method, rotates the planisphere of corresponding individual bit in opposite direction.The planisphere of rotation comprises multiple constellation point that many lines that provide with the form of vertical parallel shape or horizontal parallel shape are provided on the contrary.In operation 630, select at least two candidate constellation points for the each bit being included in in multiple bits of the symbol of transfer of data.For example, in operation 630, soft demapping is processed each in many lines that form in the planisphere for individual bit with vertical parallel shape or horizontal parallel shape, selects single candidate constellation points.

In another example, in operation 622, soft de-mapping method, not for the each bit in multiple bits, rotates in opposite direction the planisphere of corresponding individual bit, or does not comprise described rotation.Many lines that provide with the form of non-perpendicular parallel shape or non-horizontal parallel shape are provided for the each planisphere in multiple bits.In response to not rotating in opposite direction planisphere, in operation 630, soft de-mapping method is for the single candidate of each selection in many lines that provide with the form of non-perpendicular parallel shape or non-horizontal parallel shape.

In operation 640, soft de-mapping method also comprises the Euclidean distance calculating between reception signal and at least two candidates.As an example, in operation 640, calculate for the Euclidean distance between candidate and the reception signal of the each selection in many horizontal lines.In operation 650, soft de-mapping method calculates for the each LLR in multiple bits based on equation 4.Soft de-mapping method is based on individual bit b ₀logical value be distance in the candidate of " 1 " receive the nearest candidate's of signal Euclidean distance, with individual bit b ₀logical value be that distance in the candidate of " 0 " receives poor between the nearest candidate's of signal Euclidean distance, calculate for individual bit b ₀lLR.In a similar fashion, soft de-mapping method is based on individual bit b ₁logical value be Euclidean distance and the individual bit b that distance in the candidate of " 1 " receives the nearest candidate of signal ₁logical value be that distance in the candidate of " 0 " receives poor between the nearest candidate's of signal Euclidean distance, calculate for individual bit b ₁lLR.In operation 660, soft de-mapping method also comprises by preprocessor and detects bit.Each operation that the description of the Fig. 1 to Fig. 5 providing above can be applicable to describe in Fig. 6, therefore, for simplicity, will omit the description repeating.

Fig. 7 A to Fig. 7 D illustrates according to the example of the method for many line options optimal candidate of embodiment.

With reference to Fig. 7 A, the anglec of rotation of channel status and planisphere is applied to receiving the planisphere 710 of signal.The planisphere 710 that receives signal is corresponding with the rotation 16QAM planisphere 230 of the decline of Fig. 2.The planisphere 710 that receives signal is changed into original not rotation planisphere 730.For example, the state of the planisphere 710 that receives signal by the state based on I channel and Q channel be scaled changes into original rotation planisphere 720.Original rotation planisphere 720 is by being rotated to change into original not rotation planisphere 730 with the anglec of rotation of planisphere.

Change into the original not planisphere 730 of rotation in response to the planisphere 710 that receives signal, in the planisphere 710 that receives signal, correspondingly change for the projection line 712 that reception signal 711 is projected to many lines.For example, along with the planisphere 710 that receives signal is changed into original not rotation planisphere 730, projection line 712 is changed into reference line 732.Projection line 712 in the planisphere 710 of reception signal is rectangular with many lines; But the original reference line 732 not rotating in planisphere 730 is not arranged to described many lines rectangular.Selectively, the gradient of the reception signal predetermined reference line 732 based on by channel information equilibrium.

In original not rotation planisphere 730, the relative position between coordinate figure and the subpoint of the point that maintenance reception signal 731 is projected based on reference line 732 continuously.By preliminary treatment, based on such characteristic, by preliminary treatment, the expansion of for example planisphere, shrinks or rotation can be simplified.

The gradient of the projection line 712 in the planisphere 710 of reception signal is calculated as wherein, " α " represents by planisphere is rotated to the angle obtaining, " h _i" represent the state of I channel, " h _q" represent the state of Q channel.With reference to Fig. 7 B, in the situation that not there is channel degradation, in the time that the anglec of rotation of planisphere is " α " 741, the gradient of projection line 742 is tan(α).Therefore, by be illustrated in the gradient of the projection line 742 in the situation that channel degradation does not occur.

With reference to Fig. 7 C, relative when good when compare the state of Q channel with the situation of the state relative mistake of Q channel, towards the direction of the Q reference axis constellation point that distributes widely.The absolute value of the gradient of the line 751 in the relatively good situation of the state of Q channel is greater than the absolute value of the gradient of the line 753 in the situation of state relative mistake of Q channel.The gradient of the projection line 752 in the relatively good situation of the state of Q channel is less than the gradient of the projection line 754 in the situation of state relative mistake of Q channel.Therefore, the state of the gradient of projection line and Q channel is inversely proportional to.

With reference to Fig. 7 D, when compared with the situation of the state relative mistake of I channel, when the state of I channel is relatively good, towards the direction of the I reference axis constellation point that distributes widely.The absolute value of the gradient of the line 761 in the relatively good situation of the state of I channel is less than the absolute value of the gradient of the line 763 in the situation of state relative mistake of I channel.The gradient of the projection line 762 in the relatively good situation of the state of I channel is greater than the gradient of the projection line 764 in the situation of state relative mistake of I channel.Therefore, the gradient of projection line is directly proportional to the state of I channel.Therefore, the gradient of the projection line 712 in the planisphere 710 of reception signal is shown as

In the time that the planisphere 710 of reception signal is changed into original rotation planisphere 720 by channel equalization, the gradient of reference line 722 is reduced to because the state of the state of Q channel and I channel has the value between " 0 " and " 1 ", so pass through " 1/h _q" be multiplied by Q reference axis direction and by " 1/h _i" direction that is multiplied by I reference axis carries out channel equalization.By by the channel equalization rate " 1/h along I change in coordinate axis direction _i" be multiplied by with the corresponding denominator of value of the increase towards x axle and carry out channel equalization.Because serving as reasons towards the value of the increase of y axle/towards the value representation gradient of the increase of x axle, so also by by the channel equalization rate " 1/h along Q reference axis _q" be multiplied by and the corresponding molecule of value towards the axial increase of y, carry out channel equalization.Therefore by the gradient calculation of the reference line 722 in original rotation planisphere 720 be,

When original rotation planisphere 720 compensates while changing into original not rotation planisphere 730 by the anglec of rotation, with reference to the gradient calculation of line 732 be pass through with reference to the gradient of line 722 rotate in opposite direction the anglec of rotation " α " of planisphere.

The planisphere 730(using in the rotation of former beginning and end, balanced planisphere) in the reference line 732 that changes select the processing of optimal candidate as follows.Because many line comprises straight line, so represent the original many lines in planisphere 730 that do not rotate by equation 5.

[equation 5]

X=a _i(for 16QAM ,-3 ,-1,1,3)

In equation 5, " a _i" represent the coordinate of position for determining many lines.Represent to stride across by equation 6 reference line 732 that receives signal 731.

[equation 6]

$y=\frac{{\mathrm{\β}}_{\mathrm{IC}}{\mathrm{\β}}_{\mathrm{IS}}-{\mathrm{\β}}_{\mathrm{QC}}{\mathrm{\β}}_{\mathrm{QS}}}{{\mathrm{\β}}_{\mathrm{QC}}^2+{\mathrm{\β}}_{\mathrm{IS}}^2}x+(r_Q-\frac{{\mathrm{\β}}_{\mathrm{IC}}{\mathrm{\β}}_{\mathrm{IS}}-{\mathrm{\β}}_{\mathrm{QC}}{\mathrm{\β}}_{\mathrm{QS}}}{{\mathrm{\β}}_{\mathrm{QC}}^2+{\mathrm{\β}}_{\mathrm{IS}}^2}{r}_I)$

In equation 6, β _iC=h _icos, α β _iS=h _isin α, β _qC=h _qcos α, and β _qS=h _qsin α.Define the many lines intersection point crossing with reference line by equation 7.

[equation 7]

$c_I^e=a_I,c_Q^e=r_Q-(r_I-a_I)\frac{{\mathrm{\β}}_{\mathrm{IC}}{\mathrm{\β}}_{\mathrm{IS}}-{\mathrm{\β}}_{\mathrm{QC}}{\mathrm{\β}}_{\mathrm{QS}}}{{\mathrm{\β}}_{\mathrm{QC}}^2+{\mathrm{\β}}_{\mathrm{IS}}^2}$

Equation 7, for determining optimal candidate for many lines, calculates for LLR.Because in original not rotation planisphere 730, according to the spaced apart multiple constellation point that equate, so once calculate single intersection point based on equation 7, just easily calculate remaining focus based on linearity.For example, based on equation 7, the y coordinate of the intersection point on Article 1 line 733 is calculated as to " 1.4 ".In the time that the gradient of reference line 732 is assumed to be " 0.8 ", the intersection point on Article 2 line 734 is calculated as to " 1.4+2 × 0.8=0.2 ".Intersection point on Article 3 line 735 is calculated as to " 0.2+1.6=1.8 ", the intersection point on Article 4 line 736 is calculated as to " 1.8+1.6=3.4 ".As a result of, can easily realize soft de-mapping device.

Fig. 8 A illustrates bit (for example, the b for the even-numbered in the multiple bits that are included in symbol according to embodiment ₁and b ₃) calculate the example of the method for the gradient of the reference line in planisphere.Below, bit (for example, the b of the even-numbered in multiple bits ₁and b ₃) index (for example, " 1 " or " 3 ") and Odd Phases should.In addition bit (for example, the b of the odd-numbered in multiple bits, ₀and b ₂) index (for example, " 0 " or " 2 ") corresponding with even number.

With reference to Fig. 8 A, the anglec of rotation of channel status and planisphere is applied to the planisphere 810 that receives signal.The planisphere 810 that receives signal is corresponding with the rotation 16QAM planisphere 230 of the decline of Fig. 2.The planisphere 810 that receives signal can be changed into original not rotation planisphere 830.For example, the state of the planisphere 810 that receives signal by the state with I channel and Q channel be scaled changes into original rotation planisphere 820.Original rotation planisphere 820 is by being rotated to change into original not rotation planisphere 830 by the anglec of rotation of planisphere.

Because as shown in Figure 8 B, in the situation that not there is channel degradation, in the time that the anglec of rotation of planisphere is " α " 841, the gradient of projection line 842 is " tan(pi/2-α) ", is calculated as so receive the gradient of the projection line in the planisphere 810 of signal in this example, " α " represents by planisphere is rotated to the angle obtaining, " h _i" represent the state of I channel, " h _q" represent the state of Q channel.The gradient of projection line be reduced to

In the time that planisphere 810 is changed into original rotation planisphere 820 by channel equalization, with reference to the gradient calculation of line be when original rotation planisphere 820 compensates while changing into original not rotation planisphere 830 by the anglec of rotation, with reference to the gradient calculation of line be the description of previously described Fig. 7 A can be applicable to the method for the gradient of concluding reference line and uses the gradient of reference line for the method for many line options optimal candidate; Therefore, for simplicity, the description repeating will be omitted.

Fig. 9 illustrates according to the example of the soft demapping device 900 of embodiment.

With reference to Fig. 9, soft demapping device 900 comprises pretreatment unit 910, gradient calculation unit 920, candidate's selected cell 930, LLR computing unit 940, final LLR computing unit 960 and post-processing unit 970.

Pretreatment unit 910 obtains the reception signal from transmitter of the symbol that represents multiple bits.Symbol can represent M-QAM symbol, wherein, and " M "=2 ⁿand " n " is greater than two integer.

The anglec of rotation of gradient calculation unit 920 based on planisphere and channel status calculate the gradient of reference line in the planisphere for individual bit.For example, gradient calculation unit 920 calculates the gradient of projection line, and wherein, having applied in the planisphere of the anglec of rotation and channel status, described projection line is positioned to rectangular with the corresponding many lines of logical value of individual bit.Gradient calculation unit 920 carries out convergent-divergent based on channel status to the gradient of projection line, and rotation is based on the scaled straight line gradient of the anglec of rotation.Selectively, in the time that the individual bit in multiple bits belongs to first group, gradient calculation unit 920 with reference to the gradient calculation of line is in the time that the individual bit in multiple bits belongs to second group, gradient calculation unit 920 with reference to the gradient calculation of line is wherein, " h _i" represent the channel status of I channel, " h _q" represent the channel status of Q channel, " α " represents the anglec of rotation of planisphere.

The gradient of candidate's selected cell 930 based on receiving signal and reference line, from be included in the multiple constellation point the planisphere of individual bit for corresponding many single candidates of line options of logical value of described individual bit.For example, candidate's selected cell 930 comprises coordinate computing unit and constellation point selected cell.Coordinate computing unit is configured to stride across based on the gradient calculation of reference line the coordinate of the intersection point between straight line and the many lines that receive signal.Constellation point selected cell is configured to for many lines nearest constellation point of intersection point of corresponding line described in chosen distance from be included in the multiple constellation point corresponding line.

The candidate of LLR computing unit 940 based on receiving signal and selection calculates the LLR for individual bit.For example, LLR computing unit 940 calculates the Euclidean distance receiving between signal and the candidate of selection, and based on Euclidean distance, calculates the LLR for individual bit.Based on the candidate who receives signal and selection, be even number in response to the index of the individual bit in multiple bits, the LLR of the bit at the interval that it is even number that LLR computing unit 940 calculates for the index in multiple bits.Based on the candidate who receives signal and selection, be odd number in response to the index of the individual bit in multiple bits, the LLR of the bit at the interval that it is odd number that LLR computing unit 940 calculates for the index in multiple bits.

Soft demapping device 900 also comprises LUT950.Euclidean distance between LUT950 storing received signal and the candidate of selection and the candidate's of selection index.Euclidean distance for predetermined bit and the index of LLR computing unit 940 based on being stored in LUT, calculates the LLR for the bit at interval.Final LLR computing unit 960 will combine for the LLR of multiple bits to detect described multiple bit.Will be for { b ₀, b ₂..., b _n-2and { b ₁, b ₃..., b _n-1lLR combine.The bit of the LLR output decoding by post-processing unit 970 based on combination.The description of the Fig. 1 to Fig. 8 providing above can be applicable to each module shown in Fig. 9; Therefore, for simplicity, the description repeating will be omitted.

Figure 10 illustrates according to the example of the soft de-mapping method of embodiment.

With reference to Figure 10, in operation 1010, soft de-mapping method comprises that the symbol from sending from transmitter obtains or reception signal.In operation 1020, the anglec of rotation of soft de-mapping method based on planisphere and channel status calculate the gradient of reference line in the planisphere for individual bit.In operation 1030, based on the gradient that receives signal and reference line, soft de-mapping method is from being included in in the multiple constellation point the planisphere of individual bit, for the corresponding many lines of logical value of individual bit in the single candidate of each selection.In operation 1040, the candidate of soft de-mapping method based on receiving signal and selection calculates the LLR for individual bit.Each operation that the description of the Fig. 1 to Fig. 9 providing above can be applicable to describe in Figure 10; Therefore for simplicity, the description repeating will be omitted.

Unit described here can be realized with nextport hardware component NextPort.For example, nextport hardware component NextPort can comprise processor, controller, microphone, calculator, receiver, acquisition device, transmitter, amplifier, band pass filter, digital audio transducer and processing unit.Processing unit can use one or more all-purpose computer or special-purpose computer (such as, for example, processor, controller and ALU, digital signal processor, microcomputer, field programmable gate array, programmable logic cells, microprocessor or can respond and carry out in the mode of definition any other device of instruction to instruction) realize.Processing unit can operation system (OS) and one or more software application of moving on OS.Processing unit also can be in response to the execution access of software, and storage, handles, and processes and create data.For simple object, the description of processing unit is as odd number; But, it should be appreciated by those skilled in the art that processing unit can comprise multiple treatment elements and polytype treatment element.For example, processing unit can comprise multiple processors or processor and controller.In addition, different processing configuration is feasible, such as, parallel processor.

To understand in an embodiment of the present invention, although with shown in order and the operation in mode execution graph 6 and Figure 10, in the case of shown in not departing from configuration spirit and scope, can change the order of certain operations etc.According to illustrated examples, the computer program of implementing on nonvolatile computer-readable medium also can be provided, be the method for at least describing in execution graph 6 and Figure 10 by instruction encoding.

Can be by for carrying out the program command of method or one or more operation note of program command of describing at Fig. 6 and Figure 10, storing or be fixed in one or more nonvolatile computer-readable recording medium.Program command can be realized by computer.For example, computer can make processor execution of program instructions.Medium can comprise separately data file, data structure etc., or comprises data file, the data structure etc. of being combined with program command.Computer-readable medium comprises: magnetizing mediums (such as, hard disk, floppy disk and tape); Optical medium (CD ROM dish and DVD); Magnet-optical medium (such as, CD); Be specially configured for storage and the hardware unit of execution of program instructions (such as, read-only memory (ROM), random access memory (RAM), flash memory etc.).The example of program command comprises such as the machine code being produced by compiler and the file that comprises more senior code, wherein, and more senior code described in can being carried out with interpreter by computer.Program command (, software) can be distributed in the computer system of networking, thereby can store and carry out described software in the mode distributing.For example, can pass through one or more computer readable recording medium storing program for performing storing software and data.Can by the programmer in field under embodiment based on flow chart and block diagram and corresponding description thereof that the accompanying drawing that provides at this is provided, easily explain functional programs, code and code segment for realizing the embodiment disclosed herein.

Only as non-exhaustive explanation, terminal described here or device can represent to carry out with the mobile device of consistent radio communication disclosed herein or network service (such as, cell phone, PDA(Personal Digital Assistant), digital camera, portable game console, MP3 player, portable/personal multimedia player (PMP), handheld electronic book, portable PC on knee, global positioning system (GPS), flat board, transducer) and such as the device of desktop PC, high definition TV (HDTV), Disc player, Set Top Box, household electrical appliance etc.

Above, some embodiment have been described.But, should understand and can carry out various amendments.For example, if carry out described technology with different orders, if and/or be combined in a different manner assembly in described system, framework, device or circuit and/or with other assembly or its equivalent replace or replenish described in system, framework, device or circuit in assembly, can obtain suitable result.Therefore, other execution mode within the scope of the claims.

Claims (36)

1. a method of carrying out soft demapping, described method comprises:

Represent the symbol of multiple bits from receiving signal acquisition;

Constellation point from the planisphere of the each bit for multiple bits is selected at least two candidates;

Based on the Euclidean distance receiving between signal and described at least two candidates, calculate the log-likelihood ratio LLR for the each bit in multiple bits.

2. the method for claim 1, also comprises:

Calculate the Euclidean distance receiving between signal and described at least two candidates.

3. the method for claim 1, also comprises:

Use the LLR for the each bit in multiple bits, detect multiple bits.

4. the method for claim 1, wherein select at least two candidates' step to comprise:

Being the first logical value, to select at least one candidate at the each bit being included in the multiple bits in the constellation point in planisphere, is the second logical value, to select at least one another candidate at the each bit being included in the multiple bits in the constellation point in planisphere.

5. the method for claim 1, wherein select at least two candidates' step to comprise:

Corresponding with M-quadrature amplitude modulation qam symbol in response to symbol, comprise from many in the constellation point that the line of individual constellation point comprises, for every line options candidate in many lines, wherein, " M "=2 ⁿand " n " is greater than two integer.

6. method as claimed in claim 5, wherein, the constellation point in many lines is corresponding with the logical value of corresponding bits.

7. the method for claim 1, also comprises:

By the each index stores in the Euclidean distance between the reception signal of described bit and described at least two candidates and described at least two candidates in look-up table LUT.

8. method as claimed in claim 7, wherein, the step of calculating LLR comprises:

Each index in use is stored between reception signal in LUT, described bit and described at least two candidates Euclidean distance and described at least two candidates, the LLR of the bit in counting period.

9. method as claimed in claim 8, wherein, the bit at described bit and described interval is the bit of odd-numbered or the bit of even-numbered.

10. the method for claim 1, also comprises:

The planisphere that rotation receives signal is in opposite direction to provide described planisphere according to vertical parallel or horizontal parallel.

Carry out the equipment of soft demapping for 11. 1 kinds, described equipment comprises:

Pretreatment unit, is configured to represent the symbol of multiple bits from receiving signal acquisition;

Candidate's selected cell, the constellation point being configured to from the planisphere of the each bit for multiple bits is selected at least two candidates;

Log-likelihood ratio LLR computing unit, is configured to the Euclidean distance based on receiving between signal and described at least two candidates, calculates the LLR for the each bit in multiple bits.

12. equipment as claimed in claim 11, also comprise:

Metrics calculation unit, is configured to calculate the Euclidean distance receiving between signal and described at least two candidates.

13. equipment as claimed in claim 11, also comprise:

Final full LLR computing unit, is configured to use the LLR for the each bit in multiple bits, calculates the final full LLR for detection of multiple bits.

14. equipment as claimed in claim 11, wherein, candidate's selected cell is configured to: being the first logical value, to select at least one candidate at the each bit being included in the multiple bits in the constellation point in planisphere, is the second logical value, to select at least one another candidate at the each bit being included in the multiple bits in the constellation point in planisphere.

15. equipment as claimed in claim 11, wherein, candidate's selected cell is configured to: corresponding with M-quadrature amplitude modulation qam symbol in response to symbol, comprise from many in the constellation point that the line of individual constellation point comprises, for every line options candidate in many lines, wherein, " M "=2 ⁿand " n " is greater than two integer.

16. equipment as claimed in claim 11, wherein, the each index in Euclidean distance and described at least two candidates between the reception signal of bit and described at least two candidates is stored in look-up table LUT,

Wherein, LLR computing unit is configured to use the each index in Euclidean distance and described at least two candidates that are stored between reception signal in LUT, described bit and described at least two candidates, the LLR of the bit in counting period.

17. equipment as claimed in claim 16, wherein, the bit at described bit and described interval is the bit of odd-numbered or the bit of even-numbered.

18. equipment as claimed in claim 11, also comprise:

Rotary unit, is configured to rotate in opposite direction the planisphere that receives signal to provide described planisphere according to vertical parallel or horizontal parallel.

Carry out the method for soft demapping for 19. 1 kinds, described method comprises:

Based on receiving signal, in the planisphere of the bit in multiple bits of the symbol for for transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point;

Based on the candidate constellation points that receives signal and selection, calculate the log-likelihood ratio LLR of described bit.

20. methods as claimed in claim 19, wherein, corresponding with M-quadrature amplitude modulation qam symbol in response to symbol, the quantity of the candidate constellation points of selection is wherein, " M "=2 ⁿand " n " is greater than two integer.

21. methods as claimed in claim 19, wherein, symbol is with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

22. methods as claimed in claim 19, wherein, the step of calculating the LLR of described bit comprises:

Calculate the distance receiving between signal and the candidate constellation points of selection;

Based on the distance of calculating, calculate the LLR of described bit.

23. methods as claimed in claim 19, wherein, many lines comprise:

With the corresponding line of the first logical value of described bit and with the corresponding line of the second logical value of described bit.

24. methods as claimed in claim 19, wherein, corresponding with M-QAM symbol in response to symbol, every line in many lines by individual constellation point forms, wherein, and " M "=2 ⁿand " n " is greater than two integer.

Carry out the method for soft demapping for 25. 1 kinds, described method comprises:

Based on receiving signal, in the planisphere of the bit in the multiple bits that comprise for M-quadrature amplitude modulation qam symbol, select individual candidate constellation points, wherein, " M "=2 ⁿand " n " is greater than two integer;

Based on the candidate constellation points that receives signal and selection, calculate the log-likelihood ratio LLR of described bit.

26. methods as claimed in claim 25, wherein, the step of calculating the LLR of described bit comprises:

Calculating receives signal and selection distance between individual candidate constellation points;

Based on the distance of individual calculating is calculated the LLR of described bit.

27. methods as claimed in claim 25, wherein, symbol is with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

Carry out the equipment of soft demapping for 28. 1 kinds, described equipment comprises:

Selected cell, is configured to, based on receiving signal, in the planisphere of the bit in the multiple bits that comprise for M-quadrature amplitude modulation qam symbol, select individual candidate constellation points, wherein, " M "=2 ⁿand " n " is greater than two integer;

Computing unit, is configured to the candidate constellation points based on receiving signal and selection, calculates the log-likelihood ratio LLR of described bit.

Carry out the equipment of soft demapping for 29. 1 kinds, described equipment comprises:

Selected cell, is configured to based on receiving signal, in the planisphere of the bit in the multiple bits for comprising for the symbol of transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point;

Computing unit, is configured to the candidate constellation points based on receiving signal and selection, calculates the log-likelihood ratio LLR of described bit.

30. equipment as claimed in claim 29, wherein, are the bit of even-numbered in response to described bit, and the candidate constellation points that computing unit is configured to the selection that uses described bit is calculated the LLR of the bit of another even-numbered,

The bit of odd-numbered in response to described bit,

The candidate constellation points that computing unit is configured to the selection that uses described bit is calculated the LLR of the bit of another odd-numbered.

31. equipment as claimed in claim 29, wherein, computing unit is configured to calculate the distance receiving between signal and the candidate constellation points of selection, and distance based on calculating is calculated the LLR of described bit.

32. equipment as claimed in claim 29, wherein, corresponding with M-quadrature amplitude modulation qam symbol in response to symbol, wherein, the quantity of the candidate constellation points of selection is wherein, " M "=2 ⁿand " n " is greater than two integer.

33. equipment as claimed in claim 29, wherein, symbol is with to have applied the rotation qam symbol that Q postpones corresponding, corresponding with the planisphere of channel information that has reflected the channel information of I channel and the Q channel of delay for the planisphere of bit.

34. equipment as claimed in claim 29, also comprise:

Gradient calculation unit, the anglec of rotation and the channel status that are configured to based on planisphere calculate the gradient for the reference line in the planisphere of individual bit.

35. as the equipment of claim 34, wherein, selected cell is also configured to the gradient based on receiving signal and reference line, in the planisphere of the bit in the multiple bits for comprising for the symbol of transfer of data, for every line options candidate constellation points in many lines that formed by multiple constellation point.

36. equipment as claimed in claim 29, also comprise:

Full LLR computing unit, is configured to use the LLR of multiple bits to calculate the full LLR for detection of multiple bits.